Thoughts on the Semantics and Pragmatics of Rising Declaratives in English and Rise-Fall Declaratives in Hungarian

K + K = 120 / p. 247 / January 25, 2020

K + K = 120: Papers dedicated to L. Kálmán & A. Kornai on the occasion of their 60th birthdays, pp. 247–280 Thoughts on the semantics and pragmatics of rising declaratives in English and rise-fall declaratives in Hungarian

BEÁTA GYURIS Research Institute for Linguistics, Hungarian Academy of Sciences [email protected]

KEYWORDS ABSTRACT rising declarative The paper looks at the interpretation of a construction type in the Hun- rise-fall declarative garian language referred to as rise-fall declarative, which is used to en- sentence type code biased questions. Its felicity conditions are compared to those of felicity conditions rising declaratives in English, based on several recent accounts of the bias latter. It is argued that the theory proposed by Gunlogson (2003), complemented with two further conditions, can capture the licensing conditions of rise-fall declaratives in Hungarian correctly.

1. Introduction

Kálmán (2001, 101) discusses the interpretation of a type of “polar question” in Hungarian that bears a speciﬁc prosody, characterized by a rise-fall melody on all constituents in the comment part of the sentence (i.e., the part following the constituents in the topic ﬁeld):1

1 Cf. Kálmán (2001) and É. Kiss (2002) for a discussion of the syntactic structure of the Hungarian sentence. In the examples cited from Kálmán (2001) the original ĹĎ notation is retained. In the latter work, “ ” marks a rise-fall contour, and “⊥” what the authors call a “pre-tone” on unstressed deﬁnite articles and relative pronouns, which do not constitute part of the preceding character-tone. The latter phenomenon will not be discussed in what follows. The last peak within the multiple rise-fall melody falls on the penultimate syllable of the sentence, just as it does in ordinary rise-fall interrogatives, illustrated below. K + K = 120 / p. 248 / January 25, 2020

248 Beáta Gyuris

(1) ĹĎBekapcsolva hagytad a ĹĎmobiltelefont a ĹĎszínházban? switched.on left.2SG the mobile.ACC the theatre.in ‘You left the mobile phone switched on in the theatre?’ (Kálmán 2001, 101, (8))

Kálmán (2001, 101)2 argues that as compared to ordinary “polar questions”, this construction carries an interpretational surplus: “it suggests that the speaker, who presumably already knows the answer, only expects acknowledgement or explanation”.3 It is noted that this construction usually encodes echo questions uttered with a “disapproving attitude”.4 It is claimed about a different example, shown in (2), that it illustrates a special case: the aim of the question is to help find out whether some state of affairs that the conversational participants can observe in the situation was brought about by the event described in the question.5

ĹĎ ĹĎ ĹĎ (2) Felhívtad ⊥a takarítónőt, ⊥akit ajánlottam? (Azért van ilyen rend?) called.2SG the cleaner.ACC who.ACC recommended.1SG that.for be such order ‘You called the cleaner I recommended? (Is that why it looks so orderly here?)’ (Kálmán 2001, 27, (8))

Often, the speaker asks these questions to express her surprise:

ĹĎ ĹĎ ĹĎ ĹĎ (3) A Lajos berakta ⊥a kefírgombát ⊥a mélyhűtőbe? the Lajos VM.put.3SG the keﬁr.grains.ACC the freezer.into ‘Lajos put the keﬁr grains into the freezer?’ (Kálmán 2001, 27, (9))

A further illustration of the use of the same prosodic pattern is provided in (4), which appears to be an echo-question, repeating an utterance made before:6

2 This publication was the output of a seminar taught by László Kálmán at the The- oretical Linguistics Program at ELTE Budapest. He contributed signiﬁcantly to the individual chapters both in the course of the discussion and as editor of the volume. The examples cited here are from the chapters entitled Kérdések ‘Questions’ (pp. 98–135), written by Viktor Trón, and from the one entitled A topik és a kontrasztív topik ‘The topic and the contrastive topic’ (pp. 24–53), by Attila Novák and the present author. 3 “…azt sugallja, hogy a beszélő, aki már feltételezhetően tudja a választ, csak megerősítésre, magyarázatra vár”. 4 “A konstrukciót általában rosszalló visszakérdezésként használjuk.” (ibid., 101) 5 An analogous proposal about the use of the same prosody was previously made in Kálmán & Nádasdy (1994, 456). 6 The conjecture about the context is mine, the authors supply no information about contexts in general. K + K = 120 / p. 249 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 249

ĹĎ ĹĎ ĹĎ ĹĎ ĹĎ (4) Gyanútlanul ment át ⊥az utcán, ⊥amikor egyszer csak unsuspectingly went across the street.on when once only ĹĎ ĹĎ ⊥a fejére esett ⊥egy ablaktábla? the head.his.onto fell.3SG an window.pane ‘He was crossing the street unsuspectingly, when suddently a window pane fell onto his head?’ (Kálmán 2001, 27, (10))

Varga (2010, 4) also discusses questions with a similar prosodic realisation, which he characterizes as follows: “each accent in the comment is retained, and the rise-fall can appear at every accented syllable, thus forming a sequence of repeated rise-falls”. These forms encode, in his opinion, “a strongly incredulous, disbelieving yes-no question, which we ask in order to get some clariﬁcation of an unbelievable statement or experience”. His example is shown in (5). (The prosodic notation follows that of Beckman & Pierrehumbert 1986.)

(5) L* H – L% L* H – L% L* H-L% "Meghívták a "Melindát a "bulira? VM.invited.3PL the Melinda.ACC the party.onto ‘They have invited Melinda to the party? (How come?)’ (Varga 2010, 4, (3b))

Both publications cited above are primarily concerned with the prosody and the felicity conditions of the construction under consideration, and do not go into issues of formal categorization. They both refer to the relevant construction as a type of “question”.7 Our ultimate aim is to formally describe the felicity conditions of the construction type illustrated above. This paper wishes to contribute to

7 Regarding the prosodic form of the construction, an anonymous reviewer of the current paper makes the observation that the pitch can either stay in the same range throughout the whole utterance or it can drift down with each phrase that is pronounced with a rise-fall. In the reviewer’s opinion, the latter case seems to be charac- teristic of incredulous questions in general. My present suggestion is that it is not the downdrift that induces the appearance of incredulity. Instead, I assume that both phenomena are the result of information structural properties. Questions encoded by utterances containing multiple rise-falls tend to contain a lot of given material. (A question intended to ask for a speciﬁc piece of information in Hungarian can often take the form of a one-constituent utterance). One well-known reason for repeating given information is to call attention to some problem with it, e.g., the speaker’s dissatisfaction. Given information tends to be pronounced with falling pitch cross- linguistically, which might explain the downdrift in this case. Although I consider these problems worthy of further attention, I will mostly disregard the issue of the downdrift in what follows. K + K = 120 / p. 250 / January 25, 2020

250 Beáta Gyuris

this aim by looking at its formal properties, by comparing the contexts where it can appear to contexts where the construction referred to as rising declarative in Germanic and Romance languages is licensed, and by discussing to what extent the formal proposals that were put forward to capture the interpretation of the latter can be used to capture the felicity conditions of the Hungarian construction type under consideration. Section 2 ﬁrst looks at the issue of how the sentence type of this prosodically marked construction can be determined.

2. On formal category membership

The construction type illustrated in (1)–(5) is used to encode question acts, which, as a default, is done by means of interrogatives in human languages. In Hungarian there are two ways of formally marking polar interrogatives: (i) by means of the -e particle, attached to the verb as a default (“-e-interrogatives”), illustrated below in (6a), and (ii) by means of a final rise-fall intonation (“/\-interrogatives”), whose peak falls on the penultimate syllable, illustrated in (6b). At first sight, it appears to be the case that the multiple rise-fall tones constitute an additional intonational “colouring” on ordinary /\-interrogatives and encode the emotional compo- nents of incredulity, disbelief, disapproval, or the fact that the questioner asks for acknowledgement, confirmation, or for an explanation for some state of affairs. In what follows, we will argue against such an approach, and for formally characterizing the sentence type under consideration as a declarative. First, as noted in Gyuris (2017), whereas the polar interrogatives in (6a)–(6b) are perfectly grammatical if they contain negative polarity items (NPIs) like valahol is ‘anywhere’, the form with the multiple rise-fall tones is incompatible with NPIs, as shown in (6c):8

(6)a. Esik-e valahol is az eső? falls-E anywhere too the rain ‘Is it raining anywhere?’ (Gyuris 2017, 5, (7b))

8 An anonymous reviewer notes that (6c) is only ungrammatical on the non- downdrifted pronunciation of the utterance (cf. fn. 7). She/he considers all NPIs compatible with the downdrifted version of utterances bearing the multiple fall-rise prosody. Although a thorough discussion of the phenomenon has to wait for another occasion, it seems to me that the downdrifted version of (6c) should be analysed as an echoic /\-interrogative. Naturally, the validity of this suggestion can only be proven if the diﬀerent pronunciations of (6c) are considered in the appropriate contexts. K + K = 120 / p. 251 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 251

b. Esik valahol is az eső/\? ‘Is it raining anywhere?’ (ibid., 5, (7a)) c. *ĹĎEsik ĹĎvalahol is az ĹĎeső? (ibid., 6, (9))

Second, as also argued in Gyuris (2017), the Hungarian construction is incompatible with the pragmatic marker vajon ‘I wonder’, which both Ke- nesei (1992, 691) and Kálmán (2001, 99) claim to be restricted to the interrogative sentence type, and which therefore oﬀers itself as a diag- nostic property of interrogatives. The following examples illustrate the phenomenon:

(7)a. Have you been in touch with Mary lately? b. Not at all. a.′ Vajon talált-e már állást? vajon found-E already job.ACC ‘Has she already found a job, I wonder.’ a.′′ Vajon talált már állást/\? ‘Has she already found a job, I wonder.’ a.′′′ *ĹĎVajon ĹĎtalált már ĹĎállást? (Gyuris 2017, 6, (10))

Third, as shown in Gyuris (2016; 2017), negative /\-interrogatives are ambiguous between so-called “inside” and “outside” negation readings (referred to as IN and ON readings in what follows), which were ﬁrst discussed for English in Ladd (1981). (Cf. Büring & Gunlogson 2000, van Rooij & Šafářová 2013, Romero & Han 2004, Sudo 2013 for further analysis.) As also argued in Gyuris (2016; 2017), there are certain morphosyntactic features that are either only compatible with ON or only with IN readings of /\-interrogatives. If the construction type under consideration here were a /\-interrogative with an additional intonational marking, we would expect /\-interrogatives with an obligatory ON reading to have alternative pronunciations using the multiple /\-contour (abstracting away from contextual licensing conditions for a moment). This is not the case, however. (8)–(9) show that, as opposed to ordinary negative /\-interrogatives, questions encoded by the multiple rise-fall forms are not compatible with an is ‘also’ phrase, or with lack of inversion between preﬁx and verb following a negative particle nem ‘not’, which are both considered diagnostics of ON readings in Gyuris (2016; 2018): K + K = 120 / p. 252 / January 25, 2020

252 Beáta Gyuris

(8)a. Nem ment el János is moziba/\? not went VM János also movies.into ‘Didn’t John go to the movies too?’ b. *ĹĎNem ment el ĹĎJános is ĹĎmoziba?

(9)a. Nem elment moziba/\? not VM.went movies.into ‘Isn’t it the case that he went to the movies?’ b. *ĹĎNem ĹĎelment ĹĎmoziba?

Furthermore, (10a) and (10b) illustrate that whereas vala-indefinites can have both a specific and a non-specific reading in /\-interrogatives, in multiple rise-fall constructions they can only give rise to the former interpretation, just as they do in ordinary falling declaratives. (The relevant observation on the latter was made in Szabolcsi 2002, 220.)

(10)a. János nem hívott fel tegnap valakit/\? John not called VM yesterday somebody.ACC i. ‘Didn’t John call a particular person yesterday? ii. ‘Didn’t John call some person yesterday?’ (Gärtner & Gyuris 2012, 401, (25), translations amended) b. ĹĎJános ĹĎnem hívott fel ĹĎtegnap ĹĎvalakit? i. ‘John didn’t call a particular person yesterday?’ ii. *‘John didn’t call some person yesterday?’

Fourth, we can see an interesting contrast between the compatibility of /\-interrogatives versus the multiple rise-fall constructions with certain speaker-oriented adverbs. (The observations were inspired by suggestions made by Abeillé et al. 2014 about adverbials in French rising declaratives.) (11a)–(11b) show that the adverb esetleg ‘perhaps’ is grammatical in the multiple rise-fall construction, but it is ungrammatical in an ordinary /\-interrogative:

(11)a. Esetleg ĹĎbekapcsolva hagyta a ĹĎmobiltelefont a ĹĎszínházban? perhaps switched.on left.3SG the mobile.ACC the theatre.in ‘He left perhaps the mobile phone switched on in the theatre.’ K + K = 120 / p. 253 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 253

b. *Esetleg bekapcsolva hagyta a mobiltelefont a színházban/\? c. Esetleg bekapcsolva hagyta a mobiltelefont a színházban.

The falling declarative in (11c) is also compatible with esetleg, which points to a similarity between multiple rise-fall constructions and declaratives.9 As far as the adverb talán ‘perhaps’ is concerned, the situation is even more interesting: it is compatible with both structures, but it leads to a rhetorical question interpretation in the case of ordinary /\-interrogatives (which the corresponding -e-interrogatives also share):

(12)a. Talán ĹĎbekapcsolva hagyta a ĹĎmobiltelefont a ĹĎszínházban? perhaps switched.on left.3SG the mobile.ACC the theatre.in ‘He left perhaps the mobile phone switched on in the theatre?’ b. Talán bekapcsolva hagyta a mobiltelefont a színházban/\? ‘Did he perhaps leave his mobile phone switched on in the theatre?’ Intended meaning: ‘He did not leave …’

The formal and interpretational diﬀerences between /\-interrogatives and the multiple rise-fall-constructions encoding questions that we reviewed above indicate that the latter do not belong to the interrogative sentence type but to the declarative one. Therefore, in what follows, the form type under discussion here will be referred to as rise-fall declarative, abbreviated as /\-declarative. In the next section we take a closer look at an intonationally marked declarative sentence type that is referred to in Germanic and Romance languages as rising declarative. We will contrast the felicity conditions of rising declaratives, discussed in the literature, to those of Hungarian /\-declaratives, make some observations on the validity of the theoretical proposals concerning the former, and put forward a modest proposal on how the licensing conditions of the latter can be modelled formally.

9 An anonymous reviewer, while acknowledging the validity of the data in (11), notes that the negative counterpart of (11b), Esetleg nem hagyta bekapcsolva a mobiltelefont a színházban/\? is felicitous in Hungarian. Besides noting that the above structure can only give rise to an ON reading, I have no explanation for the phenomenon at the moment. K + K = 120 / p. 254 / January 25, 2020

254 Beáta Gyuris

3. Formal analyses of rising declaratives and their possible applications to /\-declaratives

3.1. Gunlogson (2003) and a proposal for extending it

Gunlogson (2003) provides a range of new observations and a formal account in terms of context update semantics regarding the use of (rising and falling) declaratives and interrogatives to encode questions in English. For the sake of brevity, in what follows, we will focus on her claims concerning the contrasts between polar interrogatives (with inversion) and rising declaratives only. The ﬁrst among them is that whereas interrogatives are generally available to ask a question in an unbiased context, declaratives are not, as (13) illustrates:

(13) At a committee hearing: a. Are you a member of the Communist party? b. #You are a member of the Communist party? (Gunlogson 2003, 1–2, (5ab))

As (14) shows, a /\-declarative is unacceptable in Hungarian in the same context, too:

(14) At a committee hearing: #Maga ĹĎtagja volt a ĹĎkommunista ĹĎpártnak? you member.its was the communist party.DAT ‘You were a member of the communist party?’

Second, similarly to ordinary interrogatives, rising declaratives do not commit the speaker to the descriptive or propositional content of the declarative. They are felicitous even if the speaker is skeptical about the truth of the latter, as the following, echoic use illustrates:

(15) A and B are looking at a co-worker’s much-dented car. A: His driving has gotten a lot better. B’s response: a. Has it? I don’t see much evidence of that. b. It has? I don’t see much evidence of that. (Gunlogson 2003, 21, (44a–b))

The next example shows that a /\-declarative is also acceptable in the same context: K + K = 120 / p. 255 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 255

(16) A and B are looking at a co-worker’s much-dented car. A: His driving has gotten a lot better. B’s response: B: Már ĹĎsokkal ĹĎjobban ĹĎvezet? Nem sok jelét látom. already much better drive.3SG not much sign.its.ACC see.1SG ‘His driving has gotten a lot better? I don’t see much evidence of that.’

The following example also illustrates lack of commitment by the speaker, but here it is not the propositional content, but the presuppositions of the interlocutor’s utterance that are challenged:

(17) A: The king of France is bald. B’s response: a. Is France a monarchy? b. France is a monarchy? (Gunlogson 2003, 2, (7a–b))

Hungarian /\-declaratives are equally ﬁne in the same context:

(18) A: The king of France is bald. B’s response: ĹĎFranciaország ĹĎkirályság? France monarchy ‘France is a monarchy?’

To account for these and analogous data, Gunlogson derives the meaning and use of rising declaratives compositionally by proposing that the declarative/interrogative form and the rising/falling intonation introduce different types of context change potentials (CCP), which are then com- bined compositionally. According to this, whereas the declarative form marks the presence of commitment to the descriptive content of the sentence, the final rise signals that this commitment is attributed to the addressee and the fall signals that it is attributed to the speaker. The formal model of the proposal uses, in addition to the concept of the Common Ground (Stalnaker 1978), the set of propositions representing the public beliefs or discourse commitments (DC) of the individual participants, referred to as DCX for participant X, and the context set associated with each discourse commitment set, referred to as csX , which consists of the set of possible worlds compatible with the propositions in DCX . Thus, the CCP of a declarative sentence is defined with respect to an individual csX , independently of the identity of X, as in (19), and the CCPs associated with rising and falling locutions as in (20)–(21), respectively, where C stands for the input and C′ for the output context: K + K = 120 / p. 256 / January 25, 2020

256 Beáta Gyuris

(19) csX + Sdecl = {w ∈ csX : the descriptive content of Sdecl is true of w} (Gunlogson 2003, 33, (74a–b))

(20) C + ↑S = C′ such that: ′ a. csAddr(C ) = csAddr(C) + S ′ b. csSpkr(C ) = csSpkr(C) (ibid., (75))

(21) C + ↓S = C′ such that: ′ a. csSpkr(C ) = csSpkr(C) + S ′ b. csAddr(C ) = csAddr (C) (ibid., (76))

Unifying the contributions of the declarative form and of the rise, the CCPs of rising declaratives look like as follows:

′ (22) C + ↑Sdecl = C such that: ′ a. csSpkr(C ) = csSpkr(C) ′ b. csAddr(C ) = csAddr(C) + Sdecl (ibid., (77))

(22) means that as a result of the utterance of a rising declarative, the context set of the speaker does not change, but that of the addressee does: only those possible worlds remain in it that are compatible with the descriptive content of Sdecl. Based on the above assumptions about the contribution of the declarative form and the rising tone, Gunlogson oﬀers the following Contextual Bias Condition on declarative questions, which proposes that rising declaratives are only compatible with contexts where the addressee is publicly committed to the proposition expressed:

(23) Contextual Bias Condition An utterance of Sdecl with descriptive content p is interpretable as a polar question in C only if csAddr(C) ⊆ p. (Gunlogson 2003, 49, (105))

Let us now consider how the above proposal accounts for the examples illustrated above. First, the theory can easily predict why echoic (a.k.a. reiterative) uses of rising declaratives, as in (15), are felicitous: the interlocutor’s previous utterance (echoed by the relevant Sdecl) explicitly indicates commitment to the descriptive content of Sdecl. Echoic uses of rising declaratives also subsume cases where the descriptive content of the Sdecl corresponds to presuppositions of the addressee’s previous utterance, as in (17), or to entailments of the proposition that the addressee has indi- cated commitment to. Both presuppositions and entailments are assumed K + K = 120 / p. 257 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 257

to automatically enter the relevant interlocutor’s cs. The following example appears problematic for this view since the proposition ‘A talked to Helena’ does not appear to be an entailment of the propositional content of A’s utterance:

(24) A: Mark and Helena are leaving for Japan this week. B: Oh … a. Did you talk to Helena? b. You talked to Helena? (Gunlogson 2003, 56, (120))

Gunlogson argues, nevertheless, that (24b) can be accounted for along the same lines as the previous examples. If p is the descriptive content of the declarative question, and q is “a relevant public commitment of the Addressee’s that serves as the basis for the inference” that p (op.cit., 58), what is required additionally for this is that q → p be accommodated as a joint commitment of the participants. Let us turn to rising declaratives that are used to encode veriﬁcation questions. Whereas in a neutral context, illustrated in (25), only interrogatives but no rising or falling declaratives seem to be felicitous in English, in a context displaying evidence for the positive answer, as in (26), all these forms are acceptable.

(25) Robin is sitting in a windowless computer room with no information about current weather conditions when another person enters. Robin says to the newcomer: a. Is it raining? b. #It’s raining? c. #It’s raining. (Gunlogson 2003, 60, (126))

(26) Robin is sitting, as before, in a windowless computer room when another person enters. The newcomer is wearing a wet raincoat and boots. Robin says: a. Is it raining? b. It’s raining? c. (I see that/So) It’s raining. (Gunlogson 2003, 61, (128))

The Hungarian counterparts of (25)–(26), illustrated in (27)–(28), respectively, show that /\-declaratives have a distribution parallel to that of rising declaratives: K + K = 120 / p. 258 / January 25, 2020

258 Beáta Gyuris

(27) Robin is sitting in a windowless computer room with no information about current weather conditions when another person enters. Robin says to the newcomer: #ĹĎEsik az ĹĎeső? falls the rain ‘It’s raining?’

(28) Robin is sitting, as before, in a windowless computer room when another person enters. The newcomer is wearing a wet raincoat and boots. Robin says: ĹĎEsik az ĹĎeső? falls the rain ‘It’s raining?’

Gunlogson accounts for the felicity of (26b) by saying that it satisﬁes the Contextual Bias Condition. There is public evidence that proposition p, de- noted by the declarative, is true, which is thus accessible to the addressee. In her opinion, however, the addressee does not base his commitment to the truth of p on this evidence but on some other information he possesses due to his position, which he came by before the evidence became available to the speaker. The role of the public evidence for the addressee is that “it enables the Addressee to recognize that the Speaker is being intentionally uninformative”, that is, that the speaker knows that the addressee knows that p is true (p. 62). The infelicity of (25b) is in turn attributed to the absence of publicly available evidence for the propositional content of the declarative. Gunlogson argues that even if Robin had access to information about the current weather conditions, unbeknownst to the newcomer, and “has good reason to be biased herself, together with the assumption that the Addressee is knowledgeable and may be presumed to have the same bias”, this would not by itself improve (25b) (p. 82). Interestingly, Gunlogson also adds the following remark to the discussion: “I want to deny that reiterative questions generally, and rising intonation speciﬁcally, are inherently associated with ‘surprise’ or ‘incredulity’, as is sometimes casually assumed.” (p. 82) There is, however, one major problem, noted by Šafářová (2007, 305), which Gunlogson’s account runs into. This concerns the apparent contradiction between the assumption that rising declaratives commit the addressee to the truth of the proposition in question and the fact that they normally still expect a response from the addressee. This problem will be addressed by the theories discussed below. Let us now consider the possibility of adopting Gunlogson’s theory for /\-declaratives in Hungarian. Echoic uses, like the one in (16), satisfy the Contextual Bias Condition (referred to as CBC in what follows). The K + K = 120 / p. 259 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 259

addressee is committed to the propositional content of his utterance, as well as to the latter’s presuppositions and entailments, thus, if the descriptive content of the echoic /\-declarative is identical to one of these, the utterance of the declarative is licensed. I want to propose, however, that in addition to the satisfaction of the CBC, there is a further necessary condition on the use of /\-declaratives, which is based on the following intuition: although the speaker must (similarly to the addressee) be committed to the publicly available evidence, she cannot have a commitment to p before the evidence became available in the context. This is based on a general requirement on question acts, captured in Searle’s Preparatory Condition 1 (Searle 1969, 66), which also plays a role in later studies of rising declaratives, discussed below. The new set of felicity conditions of /\-declaratives are shown below in (31).10 The −1 formula uses the abbreviation csSpkr(C ), which refers to the context set of the speaker in the stage of the context that preceded the one in which the /\-declarative was uttered.

(29) Felicity conditions of /\-declaratives in Hungarian (to be revised)

A /\-declarative Sdecl with propositional content p is felicitous in a context C only if

a. csAddr(C) ⊆ p. (Contextual Bias Condition, Gunlogson 2003, 49, (105)) −1 b. csSpkr(C ) * p.

(29) thus proposes that the utterance of /\-declaratives is only felicitous if they satisfy the CBC, and if the speaker was not committed to the propositional content of the /\-declarative before. As far as /\-declaratives used as verification questions, as in (27)–(28), are concerned, felicitous occurrences also obey the conditions in (29). In the case of (28), there is publicly available evidence (which the addressee is thus supposed to be committed to) that supports the truth of the propositional content p of the declarative, the speaker appears to utter the question as a reaction to this evidence, and she is not assumed to be committed to the proposition ‘It is raining’ before. In the context of (27), the first condition is not satisfied, which explains it infelicity. Let us consider now the Hungarian counterpart of (24), encoded by a /\-declarative:

10 The assumption that /\-declaratives are only licensed if the speaker is not committed independently of the evidence to the propositional content is also inspired by the deﬁnition of compelling contextual evidence in Büring & Gunlogson (2000). Cf. Gyuris (2017) for further discussion of the role of compelling contextual evidence in the licensing of polar interrogatives available for encoding question acts in Hungarian. K + K = 120 / p. 260 / January 25, 2020

260 Beáta Gyuris

(30) A: Mark and Helena are leaving for Japan this week. B: Ó, … ĹĎBeszéltél ĹĎHelénával? talked.2SG Helena.with ‘You talked to Helena?’

I believe that the acceptability of the /\-declarative in (30) can be explained if we assume that, other things being equal, whenever there is public commitment to a proposition q that entails, on the speaker’s judgment, proposition p, a /\-declarative with propositional content p is licensed. This condition can be integrated into (29) as illustrated in part (a-ii) of (31a), where BSpkr(φ) is a shorthand for ‘Speaker believes that φ’. The resulting criterion for the felicitous occurrence of /\-declaratives will be referred to as the Interlocutor Bias Condition (IBC) in what follows. This replaces the condition proposed in (29):

(31) Interlocutor Bias Condition (IBC)

A /\-declarative Sdecl with propositional content p is felicitous in a context C only if a. there is a proposition q such that (i) csAddr(C) ⊆ q, and (ii) BSpkr (q → p) −1 b. csSpkr(C ) * p.

(31) states that the necessary conditions for the licensing of a /\-declarative in a context C include the following: there should be a proposition q that the addressee is committed to (clause a-i), which, according to the speaker’s beliefs, entails p (clause a-ii), and the speaker was not previously committed to p (clause b). Let us consider how (31) accounts for the felicity of (30). At ﬁrst sight, it seems reasonable to assume that q equals the propositional content of A’s utterance, ‘Mark and Helena are leaving for Japan this week’. However, I do not think that B could be attributed a belief according to which the latter entails the proposition ‘A talked to Helena’. What the proposition ‘A talked to Helena’ can be assumed to follow from is the proposition q = ‘A made an assertion whose propositional content equals ‘Mark and Helena are leaving for Japan this week’ ’. Since this q describes a publicly available evidence, B has to be committed to it, and she can reasonably attributed the belief that q → p, on the basis of Searle’s Preparatory Condition 1 of Assertions (Searle 1969, 66), and on the basis of the fact that A observes Grice’s Communicative Principle. Note that as opposed to Gunlogson’s comments on (24), q → p does not have to be a joint belief of speaker and addressee (since the addressee does not have to have a belief that q → p), which is reﬂected in (31). K + K = 120 / p. 261 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 261

I want to turn back to echoic uses of /\-declaratives and show that they can be assimilated to veriﬁcation questions. This is based on the strategy we applied in the case of (30) above. Instead of taking the q in (31) to be the propositional content of the utterance made by the addressee, we take it to be identical to the following proposition: ‘Hearer made an utterance of an assertion/question/etc. using sentence S with descriptive content φ’. Then, (31) amounts to saying that there is a belief by the speaker that the previous utterance, based on the felicity conditions of the respective speech acts entail the propositional content of the /\-declarative.11 The last example by Gunlogson discussed here (translation of the corpus example of Beun 2000 from Dutch) is shown below in (32). Here the last question by B is considered felicitous by the author:

(32) A: Schiphol Information B: Hello, this is G.M. I have to go to Helsinki, from Amsterdam. Can you tell me which flights leave next Sunday? A: Just a moment. A: Yes, there are several flights. One leaves at 9.10, one at 11.10, and one at 17.30. B: The flight takes about three hours? (Gunlogson 2003, 58, (123))

Gunlogson claims that in spite of the fact that the CBC does not seem to be fulﬁlled in the context of (32) (since the propositional content of the rising declarative is not logically entailed by the public commitments of the addressee), the felicity of the last utterance of B can be accounted for by assuming that the rising declarative is accommodated as a question, “by making the necessary contextual adjustment to meet the Contextual Bias Condition”. Since “[t]here is no particular commitment of the Addressee’s from which the content of the declarative is taken to follow”, the author suggests that there is “a kind of blanket accommodation available for any declarative content presented by B that pertains to A’s acknowledged area of expertise, i.e., airport information”.12 As a result, “when B presents p to A declaratively, it can be taken to follow from mutual assumptions that A already knows, or is in a position to conﬁrm, p” (p. 59).

11 Note that taking q to be identical to the proposition ‘Hearer made an utterance of an assertion/question/…using sentence S with descriptive content φ’ does not inﬂuence clause (a-i) of (31), since the former describes contextual evidence, which both interlocutors, including the addressee, must be committed to. 12 According to her, the accommodation relies (i) on A being “mutually understood to be possessed of facts about some particular domain” (i.e., airport operations), and (ii) on the fact that “B has reason to believe that some proposition p is a fact and that A knows it by virtue of” (i) (i.e., “B believes p to be a fact about airport operations”). K + K = 120 / p. 262 / January 25, 2020

262 Beáta Gyuris

Independently of whether this account correctly captures the facts pertaining to English, the Hungarian counterpart of (32) containing a /\-declarative is not licensed in the relevant context:

(33) Same context as in (32) B: #ĹĎHárom ĹĎórás az ĹĎút? three hourly the way ‘The ﬂight takes three hours?’

The infelicity of (33) suggests to me that what Gunlogson refers to as the “blanket accommodation” does not apply in the case of the Hungarian construction. Clause (a) of the IBC, however, does predict that (33) is not felicitous in the context: the hearer is not publicly committed to the truth of any proposition q about which the speaker could believe that it entails the proposition ‘The ﬂight takes three hours.’ In the next section we look at Poschmann (2008), which raises several points of criticism against Gunlogson’s proposal.

3.2. Poschmann (2008)

Poschmann (2008) disagrees with the approach proposed in Gunlogson (2003), according to which declaratives used as questions uniformly involve a commitment shift from speaker to addressee. She proposes instead that the two classes of declarative questions that Gunlogson accounts for in a uniform fashion, namely, echo questions, in (15) and (17) above, and verification questions, referred to by Poschmann as confirmative questions, in (26b), should be given different treatments. Poschmann (2008, 252) argues that whereas “utterers of echo questions can easily dissociate themselves from the content of their utterance, utterers of confirmative questions obviously cannot” . The contrast can be illustrated with the help of (34) vs. (35):

(34) Echo question: A: Don’t worry. The manager has of course been informed. B: The manager has of course been informed? I wouldn’t expect that. (Poschmann 2008, 252, (9))

(35) Conﬁrmative question: At Tim’s graduation. Tim’s standing next to Sophie, a woman in her sixties. Jack to Sophie: You’re Tim’s mother? (*I don’t believe so.) (Poschmann 2008, 257, (19), context description slightly amended) K + K = 120 / p. 263 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 263

As the following examples show, judgments are similar concerning the Hungarian counterparts of the examples above:

(36) A: Ne aggódj. Az igazgatót már természetesen tájékoztatták. not worry.SUBJ.2SG the director.ACC already naturally informed.3PL ‘Don’t worry. The manager has of course been informed.’ B: Az ĹĎigazgatót ĹĎmár ĹĎtermészetesen ĹĎtájékoztatták? Nem hiszem. the director.ACC already naturally informed.3PL not believe.1SG ‘Themanager has of course beeninformed? I don’t believe it.’

(37) At Timi’s graduation. Timi’s standing next to Sophie, a woman in her sixties. Jack to Sophie: ĹĎÖn a ĹĎTimi ĹĎédesanyja? (*Nem hiszem.) you the Timi mother.her not believe.1SG ‘You’re Timi’s mother? (*I don’t believe so.)’

Poschmann proposes that “[i]n contrast to echo questions, confirmative questions seem to convey speaker commitment even though they certainly do not express the speaker’s full beliefs” (op.cit., 252). This observation is supported by the fact that confirmative questions seem to be possible both with rising and falling intonation cross-linguistically (Dutch, En- glish), whereas the intonation of echo questions is obligatorily rising. She notes, in addition, that echo-questions and echo-assertions do not necessarily attribute commitment to the addressee, they can attribute it to a third person as well, and thus suggests that the commitment shift they involve is not connected to questionhood but to their being meta-representations. Given the contrasts listed above, Poschmann (2008) argues that echo questions and confirmative questions constitute two different types of speech acts: the former involve commitment shift, and the latter speaker commitment. The speaker’s commitment “depends on the addressee’s acknowledgement: the speaker’s commitment is suspended as soon as the addressee denies the content” (op.cit., 257). After showing that a classical implicature-based theory cannot han- dle confirmative questions properly,13 Poschmann proposes an account of them that follows Zeevat (1996) and Nilsenová (2001) in assuming that “the

13 Such a theory would attribute the fact that Jack’s utterance in (35) should be interpreted as a question to an implicature: Jack’s assertion stating his assumption about Sophie being Tim’s mother would be completely uninformative for Sophie, thus it would violate the Maxim of Quantity. Thus, instead of being interpreted as an assertion, the utterance is interpreted as a question. The reason why Poschmann rejects this approach is that although the proposition that Sophie is Tim’s mother is K + K = 120 / p. 264 / January 25, 2020

264 Beáta Gyuris

speaker’s utterance does not bring an update (of the Common Ground) with the proposition p expressed by the utterance but rather with the proposition BSpkr(p) – the speaker believes that p. In order for the proposition to become part of the Common Ground (that means a commitment of both speaker and hearer), the hearer has to acknowledge it, with the update BH (p)” (Nilsenová 2001, 34). Poschmann (2008, 258) argues that whereas the context conditions are usually sufficient to trigger this acknowledgement, it is the rising intonation that explicitly realizes the call on the addressee. She proposes that “rising intonation in speech acts involving speaker-commitment can be used to ask explicitly whether the addressee acknowledges the speech act performed by the speaker”. Thus, rising intonation lends speech acts with speaker commitment a specific tentative reading, since it triggers a supplementary question about the acknowledgement (Ack) of the speech act performed (F(p)). This strategy boils down to treating confirmative questions not as questions but as complex speech acts consisting of an assertion and a question, represented formally as follows:14

(38) ASSERT (speaker, p) + QUEST (speaker, addressee, (Ack (F (p)))) (Poschmann 2008, 258, (21))

In Poschmann’s opinion, the acknowledgement of an illocutionary act involves two steps: (i) the addressee acknowledges his understanding of the speech act, and (ii) the addressee accepts or refuses the content of the speech act. Rising intonation can be used to ask for both kinds of acknowledgement, the content disambiguates as to which of them is relevant in the situation. In the case of conﬁrmative questions like (35) above, where the addressee, not the speaker, is the source of information concerning the truth of the propositional content of the utterance, acknowledgement of type (ii) plays a role. As far as “informative rising declaratives”, such as the one illustrated in (39), are concerned, acknowledgement of type (i) plays a role (given that the speaker is an authority as far as her place of origin is concerned).15

uninformative for Sophie, it is not uninformative for the Common Ground, thus the implicature does not go through after all. 14 (38) looks very similar to the interpretation Reese and Asher (2006) attribute to nuclear tag questions. The question therefore arises what predictions the present approach makes about the interpretational diﬀerence between rising declaratives and tag questions, which does exist, since the two forms cannot be replaced by each other freely in any context. 15 The example originates from Hirschberg & Ward (1995), and is also mentioned in Gunlogson (2003), but is left without speciﬁc consideration there. K + K = 120 / p. 265 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 265

(39) Informative rising declarative: Radio station DJ: Good morning, Susan. Where are you calling from? Caller: I’m from Skokie? (Poschmann 2008, 259, (23))

As far as the assertion part of the complex speech act encoded by confirmative questions according to (38) is concerned, Poschmann does not seem to assume any specific licensing conditions: the speaker commitment required for an assertion to be felicitous can either be based on contextual evidence or on the private assumptions of the speaker. The fact that the latter suffices in her opinion is demonstrated by her comments on (25), repeated below in (40). She claims that if there is a hint in the context that there was a source for a private assumption, e. g. , the internet, the rising declarative in (40b) sounds much less problematic than without it.

(40) Robin is sitting in a windowless computer room with no information about current weather conditions when another person enters. Robin says to the newcomer: a. Is it raining? b. #It’s raining? c. #It’s raining.

Although I do not want to deny that in a context modiﬁed the way sug- gested by Poschmann (40b) would be acceptable, it must be noted that this context would not meet the description given in (40) itself: Robin would not be without any information about current weather conditions. In any case, the previous discussion shows that Poschmann’s views on the acceptability of rising declaratives in a situation where only the speaker’s private information supports the truth of the propositional content are in opposition to those put forth by Gunlogson (2003). The felicity of the last utterance in (32), repeated in (41) below, ac- cords with Poschmann’s general assumptions, and is predicted in terms of (38), without having to invoke an extra mechanism as Gunlogson does: speaker B makes an assertion that is based on her private assumptions, and at the same time asks for acknowledgement about the latter’s content.

(41) A: Schiphol Information B: Hello, this is G.M. I have to go to Helsinki, from Amsterdam. Can you tell me which flights leave next Sunday? A: Just a moment. A: Yes, there are several flights. One leaves at 9.10, one at 11.10, and one at 17.30. B: The flight takes about three hours?

Although (38) seems to account well for the examples discussed here, I am not convinced that taking the utterance of rising declaratives to amount to K + K = 120 / p. 266 / January 25, 2020

266 Beáta Gyuris

the assertion of their propositional content (in addition to asking for the addressee’s opinion about it), as the formula suggests, is what the account by Nilsenová (2001), which Poschmann takes as inspiration, suggests. In particular, (13) or (41) illustrate cases where the speaker’s commitment seems to be much weaker than what is required for an appropriate assertion. Importantly, it should be noted that the Hungarian version of (39), containing a /\-declarative, is infelicitous:

(42) Radio station DJ: Good morning, Susan. Where are you calling from?

Caller: #ĹĎKarcag ĹĎmellől ĹĎtelefonálok? Karcag beside.from call.1SG ‘I’m calling from the area of Karcag?’

The fact that /\-declaratives are not licensed to appear in the context of (42) suggests that acknowledgement type (i) does not play a role in the interpretation of /\-declaratives in Hungarian. The infelicity of the /\-declarative is predicted, however, on the basis of clause (a) of the IBS: in the input context the addressee is not committed to any proposition q such that the speaker could reasonably be assumed to believe q → ‘Speaker is calling from the area of Karcag’. Interestingly, there is a form in the Hungarian language that can be used to make a speech act analogous to that in (39): this is a declarative pronounced with a ﬁnal rising tone, illustrated in (43):

(43) Caller: Karcag mellől telefonálok/. ‘I’m calling from the area of Karcag?’

The discussion of the felicity conditions of these “Hungarian rising declaratives” will, however, be left for a further occasion. Rising declaratives that are said to involve an acknowledgement of the content of the speech act in Poschmann’s framework, exempliﬁed by (13), (15), (17), and (26b) above, and those involving an acknowledgement of the form, as in (39), are also distinguished in Jeong (2017)’s framework.16 The author refers to the former type as inquisitive rising declaratives and to the latter as assertive (or informative) rising declaratives. Jeong argues that the two classes should be associated with diﬀerent felicity conditions in English. As discussed above, assertive rising declaratives do not have /\-declarative counterparts in Hungarian. There are, however, inquisitive

16 The journal article Jeong (2018) that developed from Jeong (2017) appeared too late to be given a proper discussion here. This will have to wait till another occasion. K + K = 120 / p. 267 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 267

rising declaratives, such as (41), as well, that cannot be translated into Hungarian with the help of /\-declaratives, as shown in (33). These contrasts also suggest that the felicity conditions proposed to account for the former cannot be appropriate for the analysis of the latter.17 In the next section we turn to the proposal by Gunlogson (2008), which was put forward partly as a reaction to Poschmann’s criticism, and incorporates some new insights.

3.3. Gunlogson (2008)

Gunlogson (2008) revises the author’s previous proposal, partly in order to be able to account for the data that Poschmann (2008) found incompatible with it. In order to explain the felicity of the relevant examples, Gunlogson proposes in this more recent paper that declaratives used as questions are acceptable in a context where they independently satisfy (i) felicity conditions that are associated with the use of declaratives, and (ii) conditions on the context that make the questioning interpretation possible, which is facilitated by the rising tone. In the new framework, both rising and falling declaratives are claimed to encode the speaker’s commitment, which, however, does not have to rely on contextual evidence, but can also be based on private information. The felicity conditions of initiating (that is, non-echo) uses of declarative questions rely on the concept of sourcehood:

17 (i) illustrates an inquisitive rising declarative that cannot be encoded by a /\-declarative in the same context, as (ii) shows. (i) A: The queen will arrive in five minutes. B: O.K. The manager has of course been informed? *I wouldn’t expect that. (Poschmann 2008, 252, (10)) (ii) A: The queen will arrive in five minutes. B: O.K. #Az ĹĎigazgatót ĹĎmár ĹĎtermészetesen ĹĎtájékoztatták? ‘O.K. The manager has of course been informed?’ The infelicity of (ii) comes as a surprise, since the example appears analogous to (30): if the speaker can be attributed a belief that connects the evidence and the propositional content of the /\-declarative, the utterance of the latter should be licensed in the context. Without looking deeper into possible ways of differentiating between the two contexts, I want to argue that the infelicity of (ii) can be attributed to the form of the declarative, particularly, the presence of természetesen ‘of course’, which cannot appear in utterances used to make question acts, other than those encoding echo questions. K + K = 120 / p. 268 / January 25, 2020

268 Beáta Gyuris

(44) An agent α is a source for a proposition φ in a discourse d iﬀ: a. α is committed to φ; and b. according to the discourse context, α’s commitment to φ in d does not depend on another agent’s testimony that φ in d. (Gunlogson 2008, 113, (27))

Assuming that all commitments have sources (referred to as the Source Principle in Gunlogson 2008, 117), the initiating uses of declarative questions satisfy the Rule of Initial Commitment, deﬁned as follows:

(45) Rule of Initial Commitment A speaker making a discourse commitment to φ in a context neutral with respect to φ is expected to be a source for φ. (Gunlogson 2008, 118, (39))

Gunlogson accounts for the infelicity of the declarative questions in (25), repeated again in (46), in the new framework as follows:

(46) Robin is sitting in a windowless computer room with no information about current weather conditions when another person enters. Robin says to the newcomer: a. Is it raining? b. #It’s raining? c. #It’s raining.

Using a declarative involves commitment by the speaker, which makes the speaker to be the expected source of the commitment. However, “according to what is known about Robin’s resources in the discourse situation, she is not a plausible source.” As a result, “Robin’s intention in uttering the declarative is unrecognizable, resulting in infelicity” (Gunlogson 2008, 119). The situation in (26), repeated in (47), diﬀers from the latter in that “it gives Robin a visible basis for her commitment”, and it makes it “con- ceivable in the context that Robin could reach the conclusion that it’s raining without the newcomer telling her” (idem.):

(47) Robin is sitting, as before, in a windowless computer room when another person enters. The newcomer is wearing a wet raincoat and boots. Robin says: a. Is it raining? b. It’s raining? c. (I see that/So) It’s raining.

Although the evidence that the speaker bases her commitment on is present in the dicourse context of (47), Gunlogson argues that it is not necessarily the case, as (32) or (35), repeated in (48), illustrates. “[W]hat is generally K + K = 120 / p. 269 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 269

required for felicity of a declarative is just that the discourse context allow the inference that the speaker has some basis for her choice” (ibid., 120).

(48) Conﬁrmative question: At Tim’s graduation. Tim’s standing next to Sophie, a woman in her sixties. Jack to Sophie: You’re Tim’s mother? (*I don’t believe so.)

She argues with respect to (48) that “there is no particular evidence that the woman standing next to Tim is his mother”, and “though the basis for Jack’s conjecture might be partly or entirely contextual (Sophie’s proxim- ity to Tim, say, together with the favorable odds of encountering a parent at graduation), the declaratives seem to work without requiring us to make that assumption” (Gunlogson 2008, 105). I believe, however, that without these contextual bases no speaker would be considered justiﬁed to utter the rising declarative in (48) or its Hungarian counterpart in (37), which means that their felicitous uses are covered by the CBC and the IBC, after all. The proposal according to which the speaker, who commits as source, must have adequate evidence (otherwise infelicity arises) explains only why the declarative form is felicitous in English. To account for why the declarative form can give rise to the questioning interpretation the author puts forth the following condition:

(49) Contingent Commitment Criterion An utterance of a declarative with content φ is questioning to the extent that the speaker’s commitment is understood as contingent on the addressee’s ratiﬁcation of φ. (Gunlogson 2008, 129, (48))

The role of the rising intonation is then seen by the author as marking the utterance as contingent “on some discourse condition whose identity is determined in context” (ibid., 29). Gunlogson looks at the infelicitous example (50b):

(50) (To coworker eating a piece of fruit.) a. Is that a persimmon? b. #That’s a persimmon? c. #That’s a persimmon. (Gunlogson 2008, 102, (3))

She claims that it appears reasonable to assume that “the speaker has some private basis for thinking the fruit might be a persimmon” (ibid., 131), thus the condition according to which commitments made with the help of declaratives must have sources (i.e., the Source Principle) is not K + K = 120 / p. 270 / January 25, 2020

270 Beáta Gyuris

violated. However, the Contingent Commitment Condition does seem to be violated, since there is no indication in the context that the addressee is acquainted with the name of the fruit he consumes. People are generally aware of the name of the food they eat but this follows from a generalization about people and not from properties of the context. Note that besides echoic uses of rising declaratives, the account does not apply to assertive rising declaratives, as in (39), either, since they violate the Contingent Commitment Criterion. This means that Gunlogson’s new theory is restricted to a smaller set of data than any of the previous approaches. As (51) illustrates, the Hungarian /\-declarative counterpart of (50b) is equally infelicitous in the same situation:

(51) (To coworker eating a piece of fruit.) #ĹĎEzt hívják ĹĎdatolyaszilvának? this.ACC called persimmon.DAT #‘This is called persimmon?’

Although it might follow from the common ground or from the properties of the situation that people in general or the addressee in particular knows the name of the food he is eating, it does not follow from either that the name of the fruit is a persimmon. Thus, there is no contextual evidence in the context that the addressee could be said to be committed to, and which is such that a reasonable speaker could believe it to entail the proposition ‘This is called persimmon’. As a result, the infelicity of (51) is easily accounted for in terms of clause (a) of the IBC. Consider next (52), where the appearance of a rising declarative is licensed, and its Hungarian counterpart containing a /\-declarative, in (53):

(52) (Laura has just entered the room, where Max sees her for the ﬁrst time that day.) Max: a. Did you get a haircut? b. You got a haircut? (Gunlogson 2008, 104, (8))

(53) (Laura has just entered the room, where Max sees her for the ﬁrst time that day.) Max: ĹĎLevágattad a ĹĎhajad? VM.have.cut.1SG the hair.your ‘You got a haircut?’

Gunlogson claims that the felicity of (52) is accounted for successfully on the basis of the Rule of Initial Commitment and the Contingent Com- K + K = 120 / p. 271 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 271

mitment Criterion, since the “contingency of [Max’s] commitment upon Laura’s authority is inferable in the discourse context” (op.cit., 129). I believe that the IBC also explains (53). It is without doubt that clause (b) of the IBC holds for (53): Max was not committed before seeing Laura to the truth of her having had a haircut. Let us assume that, according to clause (a) of the IBC, there is a proposition q in the context that describes the way Laura looks, which Laura herself is committed to (given that she is committed to public evidence). If the speaker can be attributed the belief that q entails the proposition ‘Laura got a haircut’, the felicity of (53) follows . The next section turns to the proposal by Malamud and Stephenson (2015).

3.4. Malamud & Stephenson (2015)

The proposal by Malamud & Stephenson (2015) regarding the interpretation of rising declaratives in English is based on two assumptions, which seem to be inspired by the theories of Gunlogson (2008) and Poschmann (2008), and thus the account seems to be the uniﬁcation of the latter two. They argue, ﬁrst, that rising declaratives introduce projected (rather than present) commitments by the speaker, which remind one of Gun- logson’s contingent commitments, and second, that these constructions add a context-dependent metalinguistic issue, which needs to be resolved, and which reminds one of Poschmann’s (2008) suggestion. Malamud and Stephenson represent the interpretation of rising declaratives in the framework proposed by Farkas and Bruce (2010), whose main features are sum- marized below. Farkas and Bruce (2010, 85) make use of a (possibly empty) set DCX for each participant X, consisting of the propositions that “X has publicly committed to during the conversation up to the relevant time, and which are not shared by all the other participants”, a set CG of propositions shared as joint discourse commitments by all participants, a stack of sentential form/meaning pairs called Table, and a set PS (“projected set”) of “projected” or “privileged” future common grounds. “The Table records what is ‘at issue’ in the conversation. When the Table is not empty, the immediate goal of the conversation is to empty it, that is, to settle the issue at hand. […] A conversation is in a stable state when its Table is empty” (ibid., 87). The system above is enriched by Malamud and Stephenson in two respects. First, they add projected commitments, “things that interlocutors are expected to become committed to in the normal course of conversation” K + K = 120 / p. 272 / January 25, 2020

272 Beáta Gyuris

(ibid., 299), which thus “represent the expected next stage of the conversation” (ibid., 288). A projected commitment of the speaker or hearer will turn into an actual commitment after the hearer has conﬁrmed it. The authors emphasize the speciﬁc nature of projected speaker commitments, “given that the speaker is always in full control of her own commitment set” (ibid., 288). Thus, “if the speaker chooses to make a projected commitment, rather than a present one, the hearer(s) can infer that the speaker has some reason to delay making a commitment that she would otherwise be willing to make” (idem.).18 Second, they add the option of introducing a metalinguistic issue, and propose that the rising tone signals the existence of such an issue. When a rising declarative is uttered, both its propositional content p and then a (possibly singleton) set of propositions MLI p is added to the Table. MLI p is a “contextually determined set of propositions, any of which would resolve the contextually determined metalinguistic issue concerning p” (Malamud & Stephenson 2015, 296). A move that simultaneously involves a commitment and a metalinguistic issue indicates to the hearer that the commitment is a projected one, pending the resolution of the metalinguistic issue. “Any aspect of the utterance’s content and form can be the subject of an MLI, as long as the speaker can give the hearer enough clues about its nature” (idem.).19 Rises, therefore, are “predicted to be possible whenever the speaker isn’t sure if a plain assertion is appropriate” (idem.). In what follows, we illustrate the procedure with some examples given by the authors, also showing the corresponding Hungarian examples encoded by /\-declaratives. In (54)–(57) the metalinguistic issue concerns the correctness of an inference by the speaker based on the interlocutor’s utterance:

(54) A and B are gossiping. A doesn’t know anything about B’s neighbor. B says, blushing, “You’ve GOT to see this picture of my new neighbor!” Without looking, A replies: A: He’s attractive? (Malamud & Stephenson 2015, 279, (2c))

(55) A: ĹĎJól néz ĹĎki? well look VM ‘He looks good?’

18 Importantly, the system including projected commitments for each participant diﬀers from Gunlogson’s contingent commitments in that Malamud and Stephenson also include projected hearer commitment, which lacks a counterpart in Gunlogson (2008). 19 This feature of the proposal was inspired by Ginzburg (1996; 2012). K + K = 120 / p. 273 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 273

(56) A and B are gossiping. A doesn’t know anything about B’s neighbor. B says, blushing, “You’ve GOT to meet my new neighbor!” A replies: A: He’s single? (Malamud & Stephenson 2015, 280, (5c))

(57) A: ĹĎNőtlen a ĹĎszomszédod? single the neighbour.your ‘Your neighbour is single?’

In (54) and (56), A infers that the neighbor is attractive or single, respectively, only indirectly; the metalinguistic issue concerns the question of whether the speaker’s inference regarding the hearer’s blushing is correct. (55) and (57) show that the corresponding Hungarian examples are felicitous in the same situations. These data would be explained by the IBC as well, if the speaker can be taken to believe that it follows from the fact that the interlocutor made the preceding utterance that the neighbour is good-looking or single, respectively. Note the similarity of these examples to (2), discussed in Kálmán (2001). The fact that the intuitions described in the latter work concerning the interpretation of the /\-declarative turn out to be very similar to our explanation of (55) and (57) suggests to me that the IBC is on the right track.20 Next, (58) illustrates a case where A is unsure about whether her opinion is called for; thus, the metalinguistic issue is whether p addresses the issue on the Table. An analogous example without a taste predicate in (59) exempliﬁes a diﬀerent kind of tentativeness, where the speaker is unsure about the speech act itself (i.e., whether the interlocutor is the right person to introduce himself to, that is, whether he is at the right place for his appointment):

(58) B hasn’t met A’s neighbor, and asks, “What do you think of your new neighbor?” A isn’t sure if B wants to know about neighborliness or suitability for dating. A replies: A: He’s attractive? (Malamud & Stephenson 2015, 280, (4c))

(59) (To a receptionist.) Hi, my name is Mark Liberman?21 (Malamud & Stephenson 2015, 281, (7))

Although I agree that the assertive rising declaratives in (58)–(59) introduce a metalinguistic issue, I believe that the speaker’s commitment to the propositional content of the rising declarative is in both cases actual,

20 I wish to thank László Kálmán for reminding me of the characterization of the interpretation of (2) in Kálmán (2001), which made me rethink the account in the version of the paper I presented at the birthday workshop. 21 Original source: Pierrehumbert & Hirschberg (1990, 290). K + K = 120 / p. 274 / January 25, 2020

274 Beáta Gyuris

rather than projected, and does not depend on the hearer’s conﬁrmation. ((39), an analogous example, illustrates the problem even better.) The Hungarian /\-declarative counterparts of these rising declaratives in (58) and (59), shown in (55) above and in (60), respectively, would not be licensed in the same situations:

(60) (To a receptionist.) #ĹĎEngem ĹĎMark ĹĎLibermannak ĹĎhívnak? I.ACC Mark Liberman.DAT call.3SG ‘My name is Mark Liberman?’

The infelicity of the Hungarian examples is expected on the basis of the IBC: contrary to what clause (a) requires, there is no proposition that the hearer is committed to and the speaker could believe to entail p. Addition- ally, in the case of (60), the speaker must also previously be committed to the relevant propositional content. The next example with a vague scalar predicate is an assertive rising declarative again, where “discourse commitments pertain to the appropriate standards of application rather than to objective facts” (Malamud & Stephenson 2015, 281):

(61) A and B are sorting paint cans in a store into a “red” bin and an “orange” bin. B points to orangishred paint and says, “What color would you say this is?” A replies: A: It’s red? (ibid., 281, (8c))

The relevant metalinguistic issue in connection with the above example is whether the standard of redness implicit in p is acceptable, given that A is not conﬁdent about her judgment. The corresponding Hungarian example in (62) below is infelicitous in the same situation, as expected, on the basis of clause (a) of the IBC:

(62) #Ez ĹĎpiros ĹĎszínű? this red coloured ‘It’s red?’

The next one is an analogous example,but without a vague scalar predicate:

(63) A teacher (B) is quizzing a student (A) on state capitals. The teacher says: “What’s the capital of New York?” The student isn’t sure of the answer, but thinks it might be Albany. The student says: It’s Albany? (Malamud & Stephenson 2015, 282, (9c)) K + K = 120 / p. 275 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 275

As expected, the /\-declarative counterpart of the example above is unacceptable:

(64) A teacher (B) is quizzing a student (A) on state capitals. The teacher says: “What’s the capital of New York?” The student isn’t sure of the answer, but thinks it might be Albany. The student says: #ĹĎNew York állam ĹĎfővárosa ĹĎAlbany? New York state capital.its Albany ‘The capital of New York state is Albany?’

Again, there is no previous commitment by the addressee that the speaker could be taken to believe that it entails the proposition ‘The capital of New York state is Albany’, thus clause (a) of the IBC is not satisﬁed. In the next section we turn to the proposal made in Farkas & Roelofsen (2017).

3.5. Farkas & Roelofsen (2017)

Farkas & Roelofsen (2017, 255) refer to the proposition that corresponds to the surface form of rising declaratives in English (as opposed to its negation) as the highlighted alternative, and propose that “both rising declaratives and tag interrogatives signal that the speaker has access to some evidence for the highlighted alternative”. In order to account for the compatibility of these two form types with particular contexts, and their incom- patibility with others, the authors suggest that the formal representation of discourse contexts proposed by Farkas and Bruce (2010), reviewed in the previous section, should be complemented, for every participant x, with a list referred to as evidence(x), which contains the possibilities “for which x has signaled to have some evidence” (idem.), and also their “credence level”, that is, “the degree to which she believes the alternative itself to be more likely than its complement” (ibid., 20). They suggest that “rising declaratives signal that the speaker’s credence in the highlighted alternative α is at most low” (ibid., 256), where low credence means that the speaker considers α to be only somewhat more likely than its negation, ¬α. As an illustration, consider the example in (65):

(65) Student: The answer to this problem is 5 because the square root of 9 is 2 and 2+3 is 5. Teacher: The square root of 9 is 2? (Farkas & Roelofsen 2017, 269, (55))

I agree with Farkas & Roelofsen (2017) in that (65) “cannot be accounted for in approaches where rising declaratives are taken to signal a ‘contingent’ or a ‘conditional’ commitment”, as in Gunlogson (2008) and Malamud & K + K = 120 / p. 276 / January 25, 2020

276 Beáta Gyuris

Stephenson (2015). A contingent or conditional commitment means that the speaker is “ready to commit to the highlighted alternative provided that her interlocutor commits to this alternative ﬁrst” (Farkas & Roelofsen 2017, 270), but in the case of (65) it does not appear to be likely that the teacher would commit to the claim under any circumstances. The authors’ own explanation sounds as follows. In the case of (65),

“the available evidence for the highlighted alternative is the student’s prior commitment to it. In this context, the teacher is assumed to be authoritative, that is, she is assumed to know whether the highlighted alternative is true or not. By her use of a rising declarative, she signals to the student that her credence in the highlighted alternative is at most low. Since she is assumed to be authoritative, this can only mean that her credence is zero, and that she is eﬀectively rejecting the student’s prior commitment and urging him to reconsider.” (op.cit., 269)

Although the part of the proposal that derives how the teacher’s use of the rising declarative can be interpreted as rejection of the highlighted alternative is rather resourceful, it remains a question why the teacher chooses the rising declarative form to formulate her question at all, instead of an ordinary positive polar interrogative form, which would unambiguously convey zero credence, and which would also be felicitous. Furthermore, I think that the proposal by Farkas and Roelofsen (2017) is contradicted by example (66), from Gunlogson (2003).

(66) A: That copier is broken. B’s response: a. Is it? Thanks, I’ll use a diﬀerent one. b. It is? Thanks, I’ll use a diﬀerent one. (Gunlogson 2003, 21, (45))

In the example above, B is not assumed to have any doubts concerning the truth of the proposition A has committed to. The example in (67) appears similar in many respects to (65), since it also involves an authority asking a question.

(67) Context: Mother sees child putting on cleats: Mother: What? You are going to play soccer? No way! You are staying home and doing your homework. (Farkas & Roelofsen 2017, 276, (68))

The Hungarian counterpart of (67), equally felicitous in the context, is shown in (68):

(68) ĹĎIndulsz ĹĎfocizni? leave.2SG play.soccer.INF ‘You are going to play soccer?’ K + K = 120 / p. 277 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 277

(68) can easily be accounted for with the help of the IBC. There is evidence that indicates the truth of the proposition q = ‘The child is putting on cleats’, which both interlocutors are committed to. If the mother believes that q, independently of any further evidence, entails the truth of p = ‘Child is going to play soccer’, the /\-declarative is licensed in the context, independently of whether the speaker wants p to become true or not. Let us ﬁnally consider the Hungarian version of (65), which is also felicitous in the relevant context:

(69) A ĹĎkilenc ĹĎgyöke ĹĎkettő? the nine square.root.its two ‘The square root of nine is two?’

I propose that (69) would be accounted for in a manner analogous to echoic uses of /\-declaratives, discussed in Section 3.1. Given the utterance of p = ‘The square root of nine is two’ by the hearer, the speaker can reasonably be assumed to believe that p is entailed, based on Preparatory Condition 1 of Assertions (Searle 1969, 66), and on the fact that the speaker considers hearer to be observing the Communicative Principle. The acceptability of (69) indicates that its licensing conditions are not influenced by whether the speaker believes in the truth of p or not. The effect of the /\-declarative is, however, influenced by the fact whether the speaker is expected to know whether p or ¬p is the case. If she is considered an expert on the issue, then the fact that she utters the /\-declarative indicates to the hearer that she doubts the truth of the propositional content. If the speaker is not supposed to know whether p is true, her question can simply be taken as asking for confirmation. The next section summarizes the results of the paper.

4. Conclusions

This paper had two major aims. On the one hand, we investigated the formal properties of the construction type in Hungarian we referred to as the /\-declarative. On the other hand, we looked at existing formal approaches to the felicity conditions of rising declaratives in English, a construction type with an apparently similar distribution, in order to identify the felicity conditions of /\-declaratives in Hungarian and consider possibilities for their formal modelling. In the course of this, we pointed out some aspects of the previous analyses that we considered problematic, but also managed to identify some distributional diﬀerences between the two constructions. K + K = 120 / p. 278 / January 25, 2020

278 Beáta Gyuris

We argued that the necessary conditions of the use of the latter should be captured by an account that is a modiﬁcation of the one proposed by Gunlogson (2003) in a way that assimilates echoic uses of the construction to conﬁrmative or inquisitive uses.

Acknowledgements

I am greatly indebted to László Kálmán for the amount and quality of work he invested in supervising my doctoral dissertation through several years. I can only hope that the present contribution approaches his high standards. I thank an anonymous reviewer and Katalin Mády for critical remarks and suggestions regarding a previous version of this work. The research was supported by the National Research, Development and Innovation Oﬃce – NKFIH, under project no. K 115922.

References

Abeillé, A., D. Godard and J.-M. Marandin. 2014. French questioning declaratives in question. In P. Hofmeister and E. Norcliﬀe (eds.) The core and the periphery. Data-driven perspectives on syntax inspired by Ivan A. Sag. Stanford, CA: CSLI Publications. Beckman, M. E. and J. B. Pierrehumbert. 1986. Intonational structure in Japanese and English. Phonology Yearbook 3. 255–309. Beun, R.-J. 2000. Context and form: Declarative or interrogative, that is the question. In H. Bunt and W. Black (eds.) Abduction, belief and context in dialogue: Studies in computational pragmatics. Amsterdam & Philadelphia: John Benjamins. 311–326. Büring, D. and C. Gunlogson. 2000. Aren’t positive and negative polar questions the same? Manuscript, UCLA & UCSC. É. Kiss, K. 2002. The syntax of Hungarian. Cambridge: Cambridge University Press. Farkas, D. and K. Bruce. 2010. On reacting to assertions and polar questions. Journal of Semantics 27. 81–118. Farkas, D. and F. Roelofsen. 2017. Division of labor in the interpretation of declaratives and interrogatives. Journal of Semantics 34. 237–289. Gärtner, H.-M. and B. Gyuris. 2012. Pragmatic markers in Hungarian: Some introductory remarks. Acta Linguistica Hungarica 59. 387–426. Ginzburg, J. 1996. Dynamics and the semantics of dialogue. In J. Seligman and D. West- erståhl (eds.) Language, logic and computation. Stanford, CA: CSLI Publications. 221–237. Ginzburg, J. 2012. The interactive stance. Oxford: Oxford University Press. Gunlogson, C. 2003. True to form: Rising and falling declaratives as questions in English. New York: Routledge. Gunlogson, C. 2008. A question of commitment. Belgian Journal of Linguistics 22. 101–136. Gyuris, B. 2016. A magyar nyelv tagadószós eldöntendő kérdő mondatainak jelentéséhez [On the interpretation of negative polar interrogatives in Hungarian]. Jelentés és Nyelvhasználat 3. 169–190. K + K = 120 / p. 279 / January 25, 2020

The semantics and pragmatics of rising declaratives and rise-fall declaratives 279

Gyuris, B. 2017. New perspectives on bias in polar questions: a study of Hungarian -e. International Review of Pragmatics 9. 1–50. Gyuris, B. 2018. On two types of polar interrogatives in Hungarian and their interac- tion with inside and outside negation. In C. Dimroth and S. Sudhoff (eds.) The grammatical realization of polarity contrast: Theoretical, empirical, and typological approaches. Amsterdam & Philadelphia: John Benjamins. 173–201. Hirschberg, J. and G. Ward. 1995. The interpretation of the high-rise question contour in English. Journal of Pragmatics 24. 407–412. Jeong, S. 2017. Deriving social meanings in an extended Lewisian model. The case of English rising declaratives. Paper presented at the workshop Questioning Speech Acts, Konstanz, 14–16 September 2017. Jeong, S. 2018. Intonation and sentence type conventions: Two types of rising declaratives. Journal of Semantics 35. 305–356. Kálmán, L. 2001. Magyar leíró nyelvtan. Mondattan I [Hungarian descriptive grammar. Syntax 1] (Segédkönyvek a nyelvészet tanulmányozásához VI [Supplementary text- books on linguistics 6]). Budapest: Tinta Könyvkiadó. Kálmán, L. and Á. Nádasdy. 1994. A hangsúly [Stress]. In F. Kiefer (ed.) Strukturális magyar nyelvtan 2: Fonológia [A structural grammar of Hungarian 2: Phonology]. Budapest: Akadémiai Kiadó. 393–467. Kenesei, I. 1992. Az alárendelt mondatok szerkezete [The structure of subordinate clauses]. In F. Kiefer (ed.) Strukturális magyar nyelvtan 1. Mondattan [A structural grammar of Hungarian 1. Syntax]. Budapest: Akadémiai Kiadó. 531–713. Ladd, D. R. 1981. A first look at the semantics and pragmatics of negative questions and tag questions. Proceedings of the Regional Meeting of the Chicago Linguistic Society contents 17. 164–171. Malamud, S. A. and T. Stephenson. 2015. Three ways to avoid commitments: Declarative force modifiers in the conversational scoreboard. Journal of Semantics 32. 275–311. Nilsenová, M. 2001. Uncertainty in common ground. ILLC Scientific Publications, MoL- 2001-03, University of Amsterdam. Pierrehumbert, J. and J. Hirschberg. 1990. The meaning of intonation contours in the interpretation of discourse. In P. R. Cohen, J. Morgan and M. E. Pollack (eds.) Plans and intentions in communication and discourse (SDF Benchmark Series in Computational Linguistics). Cambridge, MA: MIT Press. 271–311. Poschmann, C. 2008. All declarative questions are attributrive? Belgian Journal of Lin- guistics 22. 247–69. Reese, B. and N. Asher. 2006. Prosody and the interpretation of tag questions. In E. P. Waldmüller (ed.) Proceedings of Sinn und Bedeutung 11. Barcelona: Universitat Pom- peu Fabra. 393–467. Romero, M. and C.-h. Han. 2004. On negative yes/no questions. Linguistics and Philosophy 27. 609–658. Rooij, R., van and M. Šafářová. 2013. On polar questions. In R. B. Young and Y. Zhou (eds.) Proceedings of the 13th Semantics and Linguistic Theory (SALT) Conference 2013. Ithaca, NY: Linguistic Society of America and Cornell Linguistics Circle, Cor- nell University. 292–309. Šafářová, M. 2007. Nuclear rises in update semantics. In M. Aloni, A. Butler and P. Dekker (eds.) Questions in dynamic semantics. Amsterdam: Elsevier. 295–313. K + K = 120 / p. 280 / January 25, 2020

280 Beáta Gyuris

Searle, J. R. 1969. Speech acts. Cambridge: Cambridge University Press. Stalnaker, R. C. 1978. Assertion. In P. Cole (ed.) Syntax and semantics 9. New York: Academic Press. 315–332. Sudo, Y. 2013. Biased polar questions in English and Japanese. In D. Gutzmann and H.-M. Gärtner (eds.) Beyond expressives. Explorations in conventional non-truth- conditional meaning. Leiden: Brill. 277–297. Szabolcsi, A. 2002. Hungarian disjunctions and positive polarity. In I. Kenesei and P. Siptár (eds.) Approaches to Hungarian 8: Papers from the Budapest conference. Budapest: Akadémiai Kiadó. 217–239. Varga, L. 2010. Boundary tones and the lack of intermediate phrase in Hungarian (Revis- iting the Hungarian calling contour). The Even Yearbook 9. 1–27. Zeevat, H. 1996. The common ground as a dialogue parameter. In A. Benz and G. Jäger (eds.) Proceedings of MunDial 1997. Munich Workshop on Formal Semantics and Pragmatics of Dialogue. Munich: University of Munich. 195–214.